SRE Skills Bench - GMCQ: leaderboard
SRE Skills Bench — accuracy on the general multiple-choice site-reliability questions.
Metric: Accuracy (%). Source: sreskillsbench.com. Saturation forecast: Estimated already saturated. 23 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 92 | #58 |
| 2 | GPT-5.5 (High) | 91.8 | #31 (GPT-5.5) |
| 3 | Claude Opus 4.7 | 91.7 | #49 |
| 4 | GPT-5.4 | 90.3 | #79 |
| 5 | GPT-5.3 Codex | 90 | #69 |
| 6 | GPT-5.2 | 89.8 | #110 |
| 7 | GPT-5.2 Pro | 89.5 | #61 |
| 8 | GPT-5.1 | 89.2 | #135 |
| 9 | GPT-5.1 Codex Max | 89 | #100 |
| 10 | Kimi K2.5 | 88.7 | #144 |
| 11 | Claude Sonnet 4.6 | 88 | #90 |
| 12 | Claude Opus 4.5 | 87.5 | #83 |
| 13 | GPT-5 | 87 | #98 |
| 14 | Claude Opus 4.6 | 87 | #63 |
| 15 | MiniMax-M2.5 | 86.5 | #303 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=sre-skills-bench-gmcq · How It Works · Data refreshed daily, snapshot 2026-10-20.